This repository is a curated guide to open-source Machine Learning Operations (MLOps) tools, aimed at engineers and researchers seeking to understand and implement robust ML workflows. It provides a comprehensive overview of tools across various stages of the ML lifecycle, from data annotation to model deployment and monitoring, emphasizing "batteries included" solutions that are fully functional and OSI-approved.

How It Works

The guide categorizes MLOps tools based on their function within the ML pipeline, including data annotation, validation, version control, experiment tracking, model training, validation, registries, feature stores, feature engineering, deployment, serving, and monitoring. It adheres strictly to the Open Source Initiative (OSI) definition of open source, excluding tools with restrictive licenses like SSPL or those with proprietary backends.

Highlighted Details

• Comprehensive coverage of MLOps stages with tool examples.
• Strict adherence to OSI definition of open source.
• Categorization by function: data annotation, validation, version control, experiment tracking, model training, validation, registries, feature stores, feature engineering, deployment, serving, and monitoring.
• Includes links to related resources and further reading.

Maintenance & Community

This is a community-driven guide, encouraging pull requests for additions or corrections. It is maintained by Fuzzy Labs.

Licensing & Compatibility

The guide itself is not a software project but a curated list. The tools listed have various OSI-approved licenses, including Apache 2.0, MIT, GPL-3, and BSD-3-Clause. Compatibility for commercial use depends on the specific license of each tool mentioned.

Limitations & Caveats

The list is a snapshot of a rapidly evolving field and may not be exhaustive. Some tools are presented with minimal detail, requiring further investigation by the user.